| --- Log opened Sat Feb 15 00:00:59 2014 | ||
| wiking | lisitsyn: hey | 00:02 |
|---|---|---|
| thoralf | Hey wiking | 00:06 |
| -!- HeikoS [~heiko@nat-168-10.internal.eduroam.ucl.ac.uk] has joined #shogun | 00:14 | |
| -!- mode/#shogun [+o HeikoS] by ChanServ | 00:14 | |
| -!- HeikoS [~heiko@nat-168-10.internal.eduroam.ucl.ac.uk] has quit [Ping timeout: 245 seconds] | 00:18 | |
| wiking | morning :) | 00:53 |
| thoralf | Hey! | 00:53 |
| -!- thoralf [~thoralf@91-65-139-137-dynip.superkabel.de] has quit [Quit: Konversation terminated!] | 01:08 | |
| -!- bumblebee [ca4eaca2@gateway/web/freenode/ip.202.78.172.162] has joined #shogun | 06:09 | |
| -!- bumblebee is now known as Guest33104 | 06:10 | |
| Guest33104 | hey!! is anyone around? | 06:14 |
| Guest33104 | wanted to know the openings regarding opencv integration in shogun? | 06:16 |
| -!- bumblebee_ [~quassel@202.78.172.162] has joined #shogun | 06:31 | |
| bumblebee_ | ? | 06:31 |
| -!- bumblebee_ [~quassel@202.78.172.162] has quit [Client Quit] | 06:35 | |
| -!- bumblebee_ [~quassel@202.78.172.162] has joined #shogun | 06:35 | |
| @sonney2k | bumblebee_, ask pickle27 here on monday | 06:38 |
| @sonney2k | bumblebee_, and no people are usually asleep atm | 06:38 |
| bumblebee_ | k | 06:40 |
| bumblebee_ | i guess it was listed in the ideas page? Is is under maintainance ? | 06:43 |
| -!- Guest33104 [ca4eaca2@gateway/web/freenode/ip.202.78.172.162] has quit [Quit: Page closed] | 06:44 | |
| -!- dsockwell [~twgs@199.167.199.97] has left #shogun [] | 06:44 | |
| -!- bumblebee_ [~quassel@202.78.172.162] has quit [Remote host closed the connection] | 07:24 | |
| -!- pranet [73f91219@gateway/web/freenode/ip.115.249.18.25] has joined #shogun | 10:31 | |
| pranet | Hi. Am getting a "no module named shogun" error when doing "import shogun" in python. Any tips? | 10:32 |
| -!- ajgar [7aa7f31e@gateway/web/freenode/ip.122.167.243.30] has joined #shogun | 10:33 | |
| besser82 | pranet: are you using python_modular from shogun? | 10:35 |
| ajgar | I installed shogun cloning from github but I'm having trouble with the library path.. can anyone help me..? | 10:36 |
| besser82 | pranet: then you should import "modshogun" ;) | 10:36 |
| besser82 | ajgar: currently no, unfortunately. But I'll be back in ~one and a half hour to help you ;) | 10:36 |
| besser82 | ajgar: I think we can work stuff out then ;) | 10:37 |
| ajgar | besser82 no problem.. thanks a lot.. :) | 10:37 |
| pranet | ajgar. what problem with library path? | 10:41 |
| pranet | besser: umn nop. il try that | 10:41 |
| pranet | ajgar: is it the missing .so files? | 10:42 |
| ajgar | pranet: after "make all" , i did a sudo make install | 10:42 |
| ajgar | this is supposed to add libshogun to usr/lib right..? | 10:42 |
| pranet | u might have to run ldconfig explicitly. that worked for me | 10:43 |
| ajgar | ok thanks. can you pls elaborate more on that..? | 10:44 |
| pranet | have you tried compiling a cpp program yet? | 10:45 |
| ajgar | i did. compilation failed saying | 10:46 |
| ajgar | fatal error: shogun/base/init.h: No such file or directory | 10:46 |
| pranet | hmn ya i got that too. leme remember what i did | 10:47 |
| pranet | first try "ldconfig" in terminal | 10:47 |
| pranet | see if that helps | 10:48 |
| ajgar | it compiles if i explicity pass paths to lib and include folders. | 10:48 |
| ajgar | like this. g++ -I/path_to_include _L/path_to_lib test.cpp -lshogun | 10:49 |
| -!- Saurabh7 [~Saurabh7@117.235.134.71] has joined #shogun | 10:50 | |
| ajgar | pranet ldconfig did not work :( | 10:54 |
| pranet | hmn . sec | 10:55 |
| pranet | need to find out how to add the libs to default path | 10:56 |
| pranet | does ur usr/local/include have shogun? | 10:58 |
| ajgar | let me check | 10:58 |
| ajgar | no it does not. | 10:58 |
| pranet | hmn ya thats the problem. now lets see how to add it | 10:59 |
| pranet | this is how u installed right? | 11:05 |
| pranet | 1. go to the shogun repository root 2. do `mkdir build` 3. do `cmake [options] ..` (or `ccmake ..` if available). It is very recommended to use any of CMake GUIs (such as ccmake) if you feel unsure about possible parameters and configurations. 4. do `make` (and `sudo make install` if needed) | 11:05 |
| ajgar | yes | 11:05 |
| ajgar | I used these options with cmake. cmake -DCMAKE_BUILD_TYPE=Debug -DENABLE_TESTING=ON -DTRACE_MEMORY_ALLOCS=OFF -DPythonModular=ON -DBUILD_DASHBOARD_REPORTS=ON -DCMAKE_INSTALL_PREFIX="$BUILDDIR/install" | 11:05 |
| pranet | hmnn | 11:05 |
| pranet | well if u just want a working fix i could mail u the folder ur supposed to paste in usr/local/include. that may work | 11:05 |
| pranet | or just try pasting ur src/shogun folder there | 11:06 |
| ajgar | i read that copy pasting into usr/local/include could be dangerous and should not be done.. anyhow its working if I pass the path in arguments.. | 11:08 |
| ajgar | I have already spent a lot of time setting up all the dependencies. I dont want to break anything.. :D | 11:08 |
| pranet | oh didnt know about that. | 11:08 |
| pranet | well technically since the locatin is empty how could it be dangerous | 11:09 |
| pranet | worst case it doesnt work, u just delete it | 11:09 |
| pranet | again , just an opinion :P | 11:09 |
| ajgar | I am not sure how it is dangerous.. | 11:10 |
| ajgar | its working for now..let it be like this until a proper fix can be found. pls do tell me if you figure it out.. | 11:11 |
| ajgar | and thanks a lot for the help.. :) | 11:12 |
| -!- pranet_ [73f91219@gateway/web/freenode/ip.115.249.18.25] has joined #shogun | 11:13 | |
| -!- pranet [73f91219@gateway/web/freenode/ip.115.249.18.25] has quit [Ping timeout: 245 seconds] | 11:13 | |
| -!- pranet_ [73f91219@gateway/web/freenode/ip.115.249.18.25] has quit [Client Quit] | 11:14 | |
| -!- pranet [73f91219@gateway/web/freenode/ip.115.249.18.25] has joined #shogun | 11:16 | |
| pranet | npz :D | 11:20 |
| pranet | anyone knows whats python_modular? | 11:20 |
| -!- prams [~prams@122.167.243.30] has joined #shogun | 11:23 | |
| -!- ajgar [7aa7f31e@gateway/web/freenode/ip.122.167.243.30] has left #shogun [] | 11:24 | |
| -!- suryak [0e8b9b42@gateway/web/freenode/ip.14.139.155.66] has joined #shogun | 11:39 | |
| suryak | Hello there, I just had a look around the GSoC 2014 project ideas on your site and got a bit of interest to look deeply. | 11:41 |
| suryak | I have hands on experience in Machine Learning.. in which I know algorithms like logistic regression, k-means clustering, svm | 11:43 |
| suryak | and regarding programming background - I know Python, and little bit of MATLAB; I have previously done some Web programming in Django | 11:44 |
| suryak | so, basically, is there anyone I can talk to on the ideas, I appreciate; thanks | 11:44 |
| -!- pranet [73f91219@gateway/web/freenode/ip.115.249.18.25] has quit [Ping timeout: 245 seconds] | 11:47 | |
| -!- lisitsyn [~lisitsyn@80.252.20.67] has quit [Ping timeout: 252 seconds] | 11:47 | |
| suryak | anyone there? | 11:55 |
| -!- iglesiasg [~iglesiasg@524AE0A7.cm-4-3d.dynamic.ziggo.nl] has joined #shogun | 11:57 | |
| -!- mode/#shogun [+o iglesiasg] by ChanServ | 11:57 | |
| -!- suryak [0e8b9b42@gateway/web/freenode/ip.14.139.155.66] has quit [Ping timeout: 245 seconds] | 12:10 | |
| -!- prams [~prams@122.167.243.30] has quit [] | 12:17 | |
| -!- pranet [ca4eafc7@gateway/web/freenode/ip.202.78.175.199] has joined #shogun | 12:33 | |
| -!- pranet [ca4eafc7@gateway/web/freenode/ip.202.78.175.199] has quit [Quit: Page closed] | 12:46 | |
| @iglesiasg | I could really be using some help while I am building the website with the project ideas | 13:09 |
| @iglesiasg | something as simple as proof-reading could help me a lot | 13:09 |
| @iglesiasg | so please if anyone can, let me know | 13:09 |
| -!- HeikoS [~heiko@soho89-16-224-130.sohonet.co.uk] has joined #shogun | 13:19 | |
| -!- mode/#shogun [+o HeikoS] by ChanServ | 13:19 | |
| -!- Saurabh7 [~Saurabh7@117.235.134.71] has quit [Ping timeout: 248 seconds] | 13:24 | |
| @iglesiasg | hey HeikoS | 13:26 |
| -!- lisitsyn [~lisitsyn@80.252.20.67] has joined #shogun | 13:29 | |
| lisitsyn | iglesiasg: hey | 13:30 |
| @iglesiasg | lisitsyn, hello hello | 13:31 |
| lisitsyn | iglesiasg: how do you handle this stuff? | 13:33 |
| lisitsyn | repo? | 13:33 |
| @iglesiasg | lisitsyn, yes | 13:33 |
| @HeikoS | iglesiasg: hi! | 13:35 |
| lisitsyn | iglesiasg: what repo? ;) | 13:36 |
| @iglesiasg | lisitsyn, aah the one in the mail | 13:37 |
| @iglesiasg | lisitsyn, https://github.com/iglesias/shogun-gsoc2014-ideas | 13:37 |
| lisitsyn | righto | 13:37 |
| lisitsyn | iglesiasg: what do you move to the webpage right now? | 13:39 |
| @iglesiasg | lisitsyn, I just pushed the last part I have stable | 13:42 |
| @iglesiasg | lisitsyn, so pretty much, now the ideas that are really small are left, so very little work | 13:42 |
| @iglesiasg | lisitsyn, I think I can do that all right, but someone should proof read | 13:42 |
| lisitsyn | iglesiasg: let me add a few more words to kevin's opencv thing | 13:42 |
| @iglesiasg | and also, I guess the ideas that are small should be improved | 13:42 |
| @iglesiasg | lisitsyn, mmmm | 13:42 |
| lisitsyn | could you push it to the page then? | 13:42 |
| @iglesiasg | how should we do it? | 13:43 |
| @iglesiasg | soon we will start having things in two places | 13:43 |
| @iglesiasg | so I guess it is not good to update the drive more | 13:43 |
| lisitsyn | iglesiasg: well true for ideas that are online | 13:43 |
| lisitsyn | but these are not | 13:43 |
| @iglesiasg | lisitsyn, I am touching the HTML for these ones right now :) | 13:43 |
| lisitsyn | iglesiasg: ahh | 13:43 |
| @iglesiasg | lisitsyn, let me push the opencv right now for you | 13:44 |
| lisitsyn | so you have them in the html right now? | 13:44 |
| @iglesiasg | and you can go on with the html, all right? | 13:44 |
| -!- HeikoS [~heiko@soho89-16-224-130.sohonet.co.uk] has quit [Quit: Leaving.] | 13:44 | |
| lisitsyn | yeah okay | 13:44 |
| lisitsyn | lets do it | 13:44 |
| lisitsyn | iglesiasg: have you received any mails yet? | 13:45 |
| -!- HeikoS [~heiko@soho89-16-224-130.sohonet.co.uk] has joined #shogun | 13:46 | |
| -!- mode/#shogun [+o HeikoS] by ChanServ | 13:46 | |
| @iglesiasg | lisitsyn, yes | 13:47 |
| @iglesiasg | lisitsyn, and you? | 13:47 |
| lisitsyn | iglesiasg: I have one with no notion of project :D | 13:47 |
| lisitsyn | I guess this doesn't matter | 13:47 |
| lisitsyn | :D | 13:47 |
| @iglesiasg | haha | 13:48 |
| @iglesiasg | lisitsyn, ok! I just pushed everything | 13:49 |
| lisitsyn | iglesiasg: cool thanks | 13:49 |
| @iglesiasg | lisitsyn, so I have put the output kernel learning project as placeholder | 13:49 |
| @iglesiasg | lisitsyn, you can find the project you want to modify by having a look at the comments | 13:50 |
| @iglesiasg | lisitsyn, for instance, <!-- OPENCV PROJECT IDEA --> | 13:50 |
| lisitsyn | iglesiasg: alright | 13:50 |
| @iglesiasg | lisitsyn, I am working on the remaining of ML tasks now, so we try to avoid conflicts when merging later | 13:51 |
| lisitsyn | iglesiasg: PR-1 | 13:51 |
| lisitsyn | :D | 13:51 |
| -!- prams [~prams@122.167.243.30] has joined #shogun | 13:52 | |
| lisitsyn | iglesiasg: I think they should be pretty mergeable | 13:52 |
| @iglesiasg | lisitsyn, I hope so | 13:52 |
| @iglesiasg | lisitsyn, I have never really done anything with html and version control | 13:52 |
| @iglesiasg | complete surprise for me what will happen :D | 13:52 |
| -!- HeikoS [~heiko@soho89-16-224-130.sohonet.co.uk] has quit [Quit: Leaving.] | 13:53 | |
| -!- prams [~prams@122.167.243.30] has left #shogun [] | 13:53 | |
| -!- prams [~prams@122.167.243.30] has joined #shogun | 13:55 | |
| -!- pranet [ca4eafc7@gateway/web/freenode/ip.202.78.175.199] has joined #shogun | 13:58 | |
| @iglesiasg | lisitsyn, is it ok for you if I port the other ideas in infrastructure except from OpenCV one? | 14:04 |
| alcedine | I think your GSoC ideas page is a bit broken. The three ideas after "Dictionary Learning" all show the description for "Output Kernel Learning" instead. | 14:05 |
| -!- Saurabh7 [~Saurabh7@117.235.93.22] has joined #shogun | 14:06 | |
| @iglesiasg | alcedine, if you read a bit above, you will see that we are currently adding the ideas | 14:13 |
| alcedine | Ah, indeed. Sorry. (: | 14:14 |
| @iglesiasg | lisitsyn, still there? | 14:27 |
| lisitsyn | iglesiasg: yeah | 14:27 |
| @iglesiasg | lisitsyn, just opencv left to remove all placeholders | 14:28 |
| @iglesiasg | lisitsyn, maybe we could push something even if it is not completed? | 14:28 |
| lisitsyn | iglesiasg: yeah sure | 14:28 |
| @iglesiasg | lisitsyn, PR-me baby :) | 14:30 |
| lisitsyn | iglesiasg: will be back in a few mins | 14:31 |
| @iglesiasg | lisitsyn, ok, then let me push the current version in the google docs | 14:31 |
| @iglesiasg | lisitsyn, I have to go soon, and want to get rid of placeholder first | 14:31 |
| -!- Saurabh7 [~Saurabh7@117.235.93.22] has quit [Read error: Connection reset by peer] | 14:35 | |
| -!- prams [~prams@122.167.243.30] has left #shogun [] | 14:39 | |
| -!- bumblebee [~quassel@202.78.172.162] has joined #shogun | 14:42 | |
| -!- bumblebee is now known as Guest70322 | 14:42 | |
| -!- Guest70322 [~quassel@202.78.172.162] has quit [Remote host closed the connection] | 14:43 | |
| -!- bumblebee_ [~quassel@202.78.172.162] has joined #shogun | 14:44 | |
| -!- pranet [ca4eafc7@gateway/web/freenode/ip.202.78.175.199] has quit [Quit: Page closed] | 14:51 | |
| -!- Saurabh7 [~Saurabh7@117.230.130.193] has joined #shogun | 14:53 | |
| -!- Saurabh7 [~Saurabh7@117.230.130.193] has left #shogun [] | 14:53 | |
| lisitsyn | iglesiasg: ok back | 14:54 |
| @iglesiasg | lisitsyn, I think the ideas page is in good shape now | 14:56 |
| @iglesiasg | http://shogun-toolbox.org/page/Events/gsoc2014_ideas | 14:57 |
| lisitsyn | iglesiasg: looks kool indeed | 14:57 |
| @iglesiasg | it is likely that there will be errata, but we will polish those with a little bit more of time | 14:58 |
| -!- HeikoS [~heiko@soho89-16-224-130.sohonet.co.uk] has joined #shogun | 15:04 | |
| -!- mode/#shogun [+o HeikoS] by ChanServ | 15:05 | |
| -!- iglesiasg [~iglesiasg@524AE0A7.cm-4-3d.dynamic.ziggo.nl] has quit [Quit: Leaving] | 15:09 | |
| -!- votjakovr [~votjakovr@188.134.46.30] has joined #shogun | 15:24 | |
| -!- bumblebee_ [~quassel@202.78.172.162] has quit [Ping timeout: 260 seconds] | 15:27 | |
| @sonney2k | HeikoS, hey - did you find the time to write this 'how to do gsoc with us' document? | 15:54 |
| @HeikoS | sonney2k: not yet, but will start tonight | 15:55 |
| @sonney2k | HeikoS, I am just now looking at the ideas page. Looks like the creating packages & win32 port is missing and also shogun & teaching. | 15:57 |
| @sonney2k | HeikoS, I am not sure about the RBM stuff - I think a NN framework that is trained based on standard algorithms would be more valuable | 15:59 |
| @sonney2k | (standard algorithms == already available implementations) | 15:59 |
| @HeikoS | sonney2k: all this deep learning stuff is all massively depending on implementation details, so I would rather interface to existing ones | 16:00 |
| @HeikoS | yeah the packageing is important, and teaching ..... | 16:00 |
| @sonney2k | HeikoS, we have 2 labs supporting us with the teaching thing at least partially | 16:01 |
| @sonney2k | HeikoS, and this would 'just' mean write good notebooks and demos | 16:02 |
| @HeikoS | yeah | 16:02 |
| @HeikoS | its a bit hard with mentors | 16:02 |
| @HeikoS | but since this stuff is not so much about Shogun, it might be easier | 16:02 |
| * sonney2k is reading the python3.3.4 source code to understand pickling | 16:16 | |
| @HeikoS | whoooo :) | 16:22 |
| votjakovr | HeikoS: hi! how is it going with the kernel->clone() issue? I've fixed parameter's hashing framework a little bit. Now GPs are training one time. | 16:23 |
| @HeikoS | votjakovr: great work! | 16:23 |
| @HeikoS | votjakovr: yeah I looked into that, it has to do with SGVectors of CSGObjects | 16:23 |
| votjakovr | HeikoS: But it depends on kernel->clone() | 16:24 |
| @HeikoS | I think this case not not handled properly | 16:24 |
| @HeikoS | so adding a unit test for that would help (it would fail for now) | 16:24 |
| @HeikoS | and then trying to find out the issue | 16:24 |
| @HeikoS | I am not sure I even thought about that when I wrote the clone stuff (the vector of sgobjects case) | 16:24 |
| @HeikoS | votjakovr: I will get back on that later today and/or tomorrow | 16:24 |
| @HeikoS | but cool that the GPs are faster now! | 16:24 |
| @HeikoS | I am using them quite a bit for my research btw | 16:25 |
| votjakovr | HeikoS: yep, I'll send a PR after kernel->clone() problem will be fixed | 16:25 |
| @HeikoS | votjakovr: thanks! | 16:26 |
| votjakovr | HeikoS: Also i thought about adding numerical derivation to SHOGUN to find kernel derivatives wrt parameters, for which we can't do that analytically. | 16:28 |
| votjakovr | HeikoS: what do you think about that? | 16:33 |
| @HeikoS | votjakovr: I agree on that, should be general purpose and a fall back option | 16:33 |
| @HeikoS | yes, good idea1! | 16:33 |
| @HeikoS | if you want to work on that | 16:33 |
| @HeikoS | thats even a whole gsoc project, and very useful for various things | 16:33 |
| @HeikoS | in general automagic differentiation btw | 16:33 |
| @HeikoS | also non-numerical variants | 16:33 |
| @HeikoS | I gotta go now, so lets discuss later | 16:33 |
| @HeikoS | but yeah +1 from my side :) | 16:34 |
| votjakovr | HeikoS: ok, thanks :) i'll think about that | 16:34 |
| -!- votjakovr [~votjakovr@188.134.46.30] has quit [Quit: WeeChat 0.4.0] | 16:51 | |
| @sonney2k | I think torch, theano and caf* all have nice impl's written in C underneath that we might be able to use | 17:06 |
| @sonney2k | HeikoS, ^ | 17:06 |
| @sonney2k | HeikoS, also it seems a bit easier than interfacing SVMs since they don't share a common kernel | 17:07 |
| @sonney2k | HeikoS, just some weights / architecture we have to pass to them and which training algo to use | 17:07 |
| @sonney2k | HeikoS, and then fetch all the results back in | 17:07 |
| @HeikoS | sonney2k: yeah exactly, much better, and we can plug things together with the rest of shogun | 17:10 |
| @HeikoS | so no need to implement those things | 17:11 |
| @HeikoS | which is hard and messy | 17:11 |
| lisitsyn | sonney2k: sounds like my shogun vision (tm) :D | 17:32 |
| lisitsyn | just a thing that ties various things together | 17:32 |
| lisitsyn | sonney2k: what's with fastfood? is it pt.2? | 17:55 |
| -!- prams [~androirc@122.167.243.30] has joined #shogun | 17:55 | |
| -!- iglesiasg [~iglesiasg@524AE0A7.cm-4-3d.dynamic.ziggo.nl] has joined #shogun | 18:51 | |
| -!- mode/#shogun [+o iglesiasg] by ChanServ | 18:51 | |
| -!- zxtx [~zv@c-98-223-196-32.hsd1.in.comcast.net] has quit [Ping timeout: 260 seconds] | 18:52 | |
| @iglesiasg | sonney2k, is it possible to make a setup so that we don't need to use the admin interface of the website to update article pages, e.g. gsoc ideas? | 19:05 |
| -!- HeikoS [~heiko@soho89-16-224-130.sohonet.co.uk] has quit [Ping timeout: 246 seconds] | 19:11 | |
| -!- parijat [671b082a@gateway/web/freenode/ip.103.27.8.42] has joined #shogun | 19:12 | |
| parijat | sonney2k: hi, I know you guys are busy with gsoc. I had a doubt regarding PCA. Is it a good time to discuss? | 19:20 |
| parijat | sonney2k: the file was probably written by you so thats why I am reaching out to you | 19:20 |
| @iglesiasg | parijat, I am testing your file soon, let me see | 19:39 |
| parijat | iglesiasg: alright! thanks. | 19:43 |
| @iglesiasg | parijat, so there are a couple of things here | 19:44 |
| @iglesiasg | parijat, first, you should set the target dimension before calling init | 19:44 |
| parijat | iglesiasg: tried that as well. doesnt rectify the error | 19:45 |
| @iglesiasg | parijat, that will still give you a 3x4 matrix | 19:45 |
| @iglesiasg | parijat, you can check that it does different things by looking at the dimension of the matrix outputted by get_transformation_matrix | 19:46 |
| parijat | iglesiasg: yup but my target dimension is 2 right? so i am expecting 2x4 matrix | 19:46 |
| @iglesiasg | parijat, it would make sense if it does the operation in place, but I am not sure | 19:47 |
| parijat | iglesiasg: if you run kernelPCA instead with say gaussian kernel, it works perfect. I get a 2x4 matrix as output | 19:48 |
| parijat | iglesiasg: thats why i think there might be something wrong going on in PCA | 19:49 |
| @iglesiasg | parijat, in PCA::apply_to_feature_matrix for some reason it is fixed that the output has the same dimensions of the input | 19:49 |
| @iglesiasg | parijat, I don't really understand why though. I think you are right and it should output a 2x4 | 19:50 |
| parijat | iglesiasg: ya saw that! but that doesn't make sense | 19:50 |
| @iglesiasg | parijat, I mean that code is really crazy for what it actually has to do... | 19:50 |
| @iglesiasg | which is just a matrix multiplication | 19:51 |
| parijat | iglesiasg: :D | 19:51 |
| @iglesiasg | parijat, that is probably one of the things that Heiko meant when he said that (K)PCA could be pretty much rewritten | 19:51 |
| lisitsyn | iglesiasg: remember we have it in tapkee? ;) | 19:52 |
| @iglesiasg | parijat, anyhow I would say you can do like this. For your example now, work with the matrix transformation given by PCA. You can do the product in Python easily | 19:52 |
| @iglesiasg | lisitsyn, I don't actually! | 19:52 |
| parijat | iglesiasg: i tried printing the matrix m at the end of each iteration. 2 of 3 dimensions are updated per iteration. Its crazy! you should try that | 19:53 |
| parijat | iglesiasg: ok alright! let me try that | 19:53 |
| @iglesiasg | parijat, I used PCA for LMNN just using the outputted matrix transformation, so that should be ok | 19:55 |
| parijat | iglesiasg: alright I will have a look at your code for reference as well. | 19:56 |
| parijat | iglesiasg: but we should also try fixing this PCA bug (if there is one) sometime soon :) | 19:57 |
| @iglesiasg | parijat, yeah sure. It would be nice to remove the operations that use cblas there and use eigen to start with | 19:57 |
| @iglesiasg | parijat, take it as another entrance related to the fundamental ML project if you want ;) | 19:58 |
| parijat | iglesiasg: :D ya sure! my pleasure. | 19:58 |
| parijat | iglesiasg: thanks for looking into this. | 19:59 |
| @iglesiasg | no problem! | 20:00 |
| @iglesiasg | parijat, I am just going to answer the mail in the mailing list so there is an answer there | 20:10 |
| parijat | iglesiasg: ya sure go ahead. I was just going through the pca transform used in LMNN, I am mostly just translating the code to python :) | 20:14 |
| -!- HeikoS [~heiko@soho89-16-224-130.sohonet.co.uk] has joined #shogun | 20:15 | |
| -!- mode/#shogun [+o HeikoS] by ChanServ | 20:15 | |
| @iglesiasg | parijat, all right! Probably you just need to do get_transformation_matrix to get the PCA matrix and then numpy.dot to perform matrix multiplication between the feature matrix and the transformation matrix | 20:20 |
| -!- zxtx [~zv@c-98-223-196-32.hsd1.in.comcast.net] has joined #shogun | 20:20 | |
| parijat | iglesiasg: ya, but before that subtracting out the mean is also necessary maybe? | 20:21 |
| @iglesiasg | parijat, that is an additional step you may or may not do | 20:22 |
| @iglesiasg | parijat, I see that the PCA code does it, but it is not a compulsory step in standard PCA (I believe) | 20:22 |
| parijat | iglesiasg: ya thats why I planned on doing it. Maybe I will do it anyways. It isn't much of a work | 20:23 |
| @iglesiasg | indeed, just another 2-3 lines at most it must be :) | 20:24 |
| -!- parijat [671b082a@gateway/web/freenode/ip.103.27.8.42] has quit [Ping timeout: 245 seconds] | 20:36 | |
| -!- huska1 [~huska@poutine.molgen.mpg.de] has quit [Quit: WeeChat 0.4.3] | 20:38 | |
| lisitsyn | iglesiasg: I think mean is important here ;) | 20:41 |
| @iglesiasg | lisitsyn, I saw tapkee has it too so I started wondering hehe | 20:43 |
| lisitsyn | iglesiasg: I think mean is changing the angle | 20:43 |
| lisitsyn | speaking of geometry | 20:43 |
| -!- huska [~huska@poutine.molgen.mpg.de] has joined #shogun | 20:43 | |
| lisitsyn | any prinicipal axis has different angle when you don't subtract mean and when you do | 20:44 |
| -!- prams [~androirc@122.167.243.30] has quit [Ping timeout: 252 seconds] | 20:48 | |
| -!- lisitsyn [~lisitsyn@80.252.20.67] has quit [Read error: Connection reset by peer] | 20:50 | |
| -!- lisitsyn [~lisitsyn@80.252.20.67] has joined #shogun | 20:50 | |
| -!- iglesiasg [~iglesiasg@524AE0A7.cm-4-3d.dynamic.ziggo.nl] has quit [Ping timeout: 252 seconds] | 20:54 | |
| -!- iglesiasg [~iglesiasg@524AE0A7.cm-4-3d.dynamic.ziggo.nl] has joined #shogun | 21:07 | |
| -!- mode/#shogun [+o iglesiasg] by ChanServ | 21:07 | |
| -!- iglesiasg [~iglesiasg@524AE0A7.cm-4-3d.dynamic.ziggo.nl] has quit [Client Quit] | 21:07 | |
| -!- zxtx [~zv@c-98-223-196-32.hsd1.in.comcast.net] has quit [Ping timeout: 245 seconds] | 22:05 | |
| -!- zxtx [~zv@c-98-223-196-32.hsd1.in.comcast.net] has joined #shogun | 22:19 | |
| -!- prams [~androirc@122.167.243.30] has joined #shogun | 22:28 | |
| -!- thoralf [~thoralf@91-66-33-38-dynip.superkabel.de] has joined #shogun | 23:35 | |
| thoralf | Hey. | 23:35 |
| -!- thoralf [~thoralf@91-66-33-38-dynip.superkabel.de] has quit [Quit: Konversation terminated!] | 23:44 | |
| --- Log closed Sun Feb 16 00:00:01 2014 | ||
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!